On a winter Saturday each month, a small group of volunteers gathers on Bibra Drive in Perth's southern suburbs to plant seedlings into a strip of reclaimed bushland running between Bibra Lake and North Lake. The ground under their boots was, until early 2017, a partly bulldozed road corridor — the start of Roe Highway Stage 8, the most fought-over piece of infrastructure Western Australia has tried to build this century. Roughly 18 hectares (approx.) of native vegetation were cleared here over the summer of 2016-17 before construction was suspended; the volunteers, working under a ten-year (2018-2028) rehabilitation program run by the Department of Biodiversity, Conservation and Attractions with the City of Cockburn, have so far put more than 270,000 seedlings back into that scar (approx., as reported; Rehabilitating Roe 8 project, City of Cockburn). A few kilometres north-west, container trucks still work their way along Fremantle's suburban streets to reach the working harbour Roe 8 was originally built to serve. Cancelling the road answered the wetlands' problem. It did not answer the port's.
What the government did
Roe 8 was the first stage of the Perth Freight Link, a roughly $1.9 billion (approx.) project announced by the Barnett Liberal-National government in May 2014 to build a continuous freight corridor from Kewdale to Fremantle Harbour: a five-kilometre extension of Roe Highway to Stock Road (Roe 8), a further stage bypassing fourteen sets of traffic lights to Stirling Highway (Roe 9), and a final connection into the port itself (Infrastructure Australia, Perth Freight Link project evaluation). The Commonwealth government offered $1.2 billion toward the project — money recorded in the federal budget as a standing commitment to whichever WA government built it (WA Parliament, Legislative Assembly Hansard, 25 June 2019). The environmental approval process was contested from the start. In February 2013 the Aboriginal Cultural Material Committee recommended against granting consent for the road, citing the North Lake-Bibra Lake corridor's significance to Noongar people — the area holds sites linked to the Waugal, the creation serpent of Noongar tradition, and has recorded Aboriginal occupation stretching back more than 5,000 years. Three months later the Department of Aboriginal Affairs changed its guidelines for assessing sacred sites under section 5(b) of the Aboriginal Heritage Act, and consent proceeded regardless. The EPA separately assessed and approved the Roe 8 proposal, with the Environment Minister confirming that approval on 2 July 2015. Save Beeliar Wetlands Inc took the EPA to the WA Supreme Court, and on 16 December 2015 Chief Justice Wayne Martin found the EPA's assessment invalid because it had failed to properly apply its own published policy. That ruling was unanimously overturned by the Court of Appeal on 15 July 2016, and the High Court declined to disturb that outcome on 16 December 2016 — meaning that, by the time of the March 2017 state election, Roe 8 had cleared every court test put to it (Supreme Court of Western Australia, Save Beeliar Wetlands (Inc) v Jacob, judgment summary, 16 Dec 2015). A construction alliance led by CPB Contractors, with Georgiou, WA Limestone, GHD, AECOM and BG&E, was awarded the head contract in October 2016, and clearing began that December. Opposition escalated into direct blockades; on 12 January 2017 an "Enough is Enough" protest drew roughly 1,000-2,000 people to the clearing site at Bibra Lake, and close to 150 arrests were recorded across the three-month campaign of daily protests that followed, prompting an independent review into the policing response.
The March 2017 election was fought partly on this single road. Labor, led by Mark McGowan, campaigned explicitly on scrapping Roe 8; Labor won in a landslide, and on 12 March 2017 Main Roads and the construction alliance agreed to suspend all work. The state renegotiated the alliance's contract into three unrelated projects — duplicating Wanneroo Road between Joondalup Drive and Flynn Drive (about $31 million), building the Murdoch Drive-Kwinana Freeway-Roe Highway access upgrade serving Fiona Stanley Hospital (about $100 million) and duplicating Armadale Road between Anstey Road and Tapper Road (about $145 million) — together worth roughly $276 million (approx.), funded from money originally earmarked for Perth Freight Link. The larger share of the Commonwealth's $1.2 billion commitment sat unresolved for years: Prime Minister Malcolm Turnbull initially insisted it could not be redirected to other WA projects, and the provision remained on the federal budget until the Albanese government formally removed it in the 2022-23 Federal Budget — roughly five years after WA had cancelled the road it was meant to fund. Under Premier Roger Cook, the state took the final legal step in the saga: on 17 January 2025, Planning Minister John Carey finalised a Metropolitan Region Scheme amendment removing the Primary Regional Roads reservation over the Roe 8/9 corridor, rezoning its environmentally and culturally significant sections as Parks and Recreation and the remaining cleared land for urban uses — a change the Cook government describes as making the road legally impossible to revive, after a last-day disallowance attempt by the Greens failed to stop it (WA Government media statement, "Significant amendment winds up Roe Highway extension", 17 Jan 2025). In parallel, the Cook government has pursued a different answer to the underlying freight problem: relocating the container port itself, through the Westport program, rather than building a new road into the old one.
The strongest case
An irreplaceable wetland, saved on an explicit mandate
Put the case for cancellation at its strongest, and it does not rest on the courts — it rests on democracy and on what would have been permanently lost. Every court that examined the EPA's approval of Roe 8, up to the High Court, ultimately let it stand; cancellation cannot be defended as the law forcing the government's hand. But the 2017 election put a clear, specific choice to WA voters — scrap this road or build it — and Labor's landslide win is about as direct a mandate as state politics produces. What stood to be destroyed if the road had gone ahead was not recoverable: a wetland system with more than 5,000 years of recorded Noongar occupation, sites connected to the Waugal creation story, and habitat for the endangered Carnaby's black cockatoo, all inside a growing city that will not get another such site back once it is gone. The Aboriginal Cultural Material Committee's own 2013 recommendation against the road, made before the assessment guidelines that overrode it were quietly changed, is a serious mark against how thoroughly the original approval considered what was actually at stake. Measured against roughly $276 million in contract value that was redeployed into three other, genuinely needed road upgrades rather than wasted outright, the fiscal cost of stopping Roe 8 was real but bounded — and bounded, recoverable costs are a defensible trade against an unbounded, irreversible loss.
The evidence
The freight problem Roe 8 was meant to solve has not gone away; it has simply been left for a different, larger project to fix. Fremantle Port remains a working container harbour hemmed in by the residential suburbs of Fremantle and North Fremantle, moving cargo without the dedicated freight corridor Roe 9 would have built. Meanwhile the port keeps setting records: Fremantle Ports' own trade reporting shows container throughput reaching roughly 892,000 TEU (twenty-foot equivalent units, the standard measure of container volume) in 2024-25 (approx.), the second consecutive record year, against a stated capacity ceiling of around 1.4 million TEU in the Westport business case (Fremantle Ports, Trade Reports; WA Government, Westport Business Case summary, 11 Nov 2024).
The government's own Westport business case, released on 11 November 2024, put the cost of a new Kwinana container port at roughly $7.2 billion in government investment, against an estimated $244 billion (approx.) economic cost to WA over coming decades — about $5 billion a year — if the state fails to act before Fremantle's capacity runs out (WA Government, Westport Business Case summary, 11 Nov 2024). But the date that ceiling will actually be reached has moved substantially depending on when you ask. State planning around 2020 canvassed a new port being needed by 2032; the 2024 business case pushed the point at which Fremantle's capacity is exhausted out to around 2038-2040, or the mid-2030s under higher-growth scenarios; by May 2026, Parliamentary Secretary for Ports Lisa O'Malley told a Budget Estimates hearing that Fremantle's existing capacity was now not expected to be reached until somewhere between 2045 and 2050 (as reported; WA Legislative Council, Budget Estimates hearings, May 2026, cited in Sound Telegraph reporting, 21 May 2026). Genuine trade forecasting does get revised as new data arrives, and a longer runway is not inherently bad news. But a target date that has moved by well over a decade across three sets of official statements is also the kind of pattern the WA Auditor-General has flagged elsewhere in the state's major-project pipeline as a sign that initial advice to government is not capturing the real picture.
The critique
Two separate failures sit inside this one story, and they should not be collapsed into each other. The first is the original approval process: a heritage committee's 2013 recommendation against the road was effectively neutralised by a guideline change three months later, and the EPA's own assessment was found by the state's Chief Justice to have ignored its published policy — even though that finding was later overturned on appeal. A process that needs a guideline rewritten to get past its own heritage advisers, and whose environmental assessment is found wanting by the trial judge who examined it most closely, is not a process that inspires confidence in how the original approval was reached, whatever the appeal courts ultimately concluded about its legality.
The second failure belongs to what came after cancellation, and it is squarely one this masthead can put to the continuing WA Labor government, from McGowan through to Cook. Stopping a bad road was not the same as solving the freight-access problem that road was built to address, and on the record above, that problem has now sat substantially unresolved for close to a decade. Roe 8 was suspended in March 2017. The Westport alternative was not formally recommended to government until the November 2024 business case — seven and a half years later — and by mid-2026 it had reached only a planning-and-definition footing: $273 million committed for design, approvals and land acquisition, a further $68.4 million allocated across 2025-26 to 2029-30 for the Westport Office and early works, and a Westport Authority Bill introduced to Parliament on 10 March 2026 but not yet passed. No final investment decision has been made. The Commonwealth's contribution to date is $33.5 million — roughly half of one per cent of the $7.2 billion the project is expected to need — a far smaller down payment than the $1.2 billion Canberra had once put on the table for Roe 8/9. Meanwhile Fremantle Port has posted back-to-back record container years, narrowing whatever margin remains before the capacity ceiling — wherever it actually sits — is reached. A government does not have to be blamed for the geography that put a 19th-century harbour inside a 21st-century city. It can fairly be asked why, nine years after correctly identifying that Roe 8 was the wrong answer, its replacement still has no funded construction start, no locked-in delivery date and no passed enabling legislation.
Cancelling Roe 8 answered one question in 2017. Nine years on, Fremantle Port is still waiting for the other one.
The risk to the future
A ten-year gap the state cannot plant its way out of
The nearer risk is capacity itself. Fremantle Port's throughput has set a new record in each of the last two reported years, and the business case's own modelling puts the cost of failing to act before capacity is exhausted at roughly $5 billion a year (approx.) in lost economic activity for the state. Every year Westport's final investment decision slips is a year taken directly out of whatever buffer remains between now and that ceiling — a buffer whose own estimated date has already moved by more than a decade across three official statements since 2020.
The second risk is that Westport repeats the pattern the WA Auditor-General has already documented in the state's other mega-project, METRONET: a headline cost agreed years before delivery, on a program whose own timeline language has proven elastic. A $7.2 billion (approx.) figure fixed in a 2024 business case, attached to a project that has not yet reached a final investment decision or passed its enabling legislation, carries the same structural exposure to cost and schedule growth that has already shown up elsewhere in WA's infrastructure pipeline. Having taken the harder, slower, more defensible road on Beeliar Wetlands, the state now needs to show it can move faster and more predictably on the port that decision was supposed to make possible.